But I felt that it was too specialized of a case.

I found that if the PC side is running a program that has multiple threads (the usb not being the highest priority, but the pci bus being the higher priority) and is not sitting on the host controller waiting for records that the host controller will continue to capture data from the device and will actually overwrite its own buffer regardless of which mode of data transmission. This, of course, you have no control over how the hardware acts. So my only patch and not really a solution was to add a record counter in which you can check the counter to see if you have missed any records that needed to be paired with separate bus data that may be present and needs to be thrown away. This is an example of an extremely high speed waveform capturing device located in the PCI bus matched with encoded data from the USB bus as being a highly asynchronous system.

That being said, the software provided will show you that you have not missed a single record, but in fact your PC program that you developed will tell another story.

If there is a better patch/solution and even concept, I would be glad to hear it. The best solution I have thought of to date was to actually create the high speed waveform capturing card and integrate the encoded portion along with it, but as it stands the waveform capturing is currently bought.